ESS - Netzbezug trotz vollem Akku

Something similar had me scratching my head for a couple of weeks last winter with my van-conversion setup — though obviously on a smaller scale than a three-phase Multiplus-II arrangement.

I was running a Victron MultiPlus-II 12/3000 alongside a Fogstar Drift 12V 200Ah LiFePO4 and noticed the system was still pulling from shore power even when the battery showed well above 90% SOC. Turned out the issue was a mismatch between what the JK BMS was reporting and what the MultiPlus was actually seeing via the DVCC settings in VenusOS. The SOC read fine on the Cerbo GX display but the charge parameters weren't being respected properly.

For anyone running a DIY battery — especially with a JK BMS — it's worth double-checking:

• DVCC is enabled and the BMS is the selected charge current controller
• Your CVL, CCL and DCL values are actually being passed through correctly (you can verify in the VRM portal under Device List → Battery Monitor)
• The ESS mode is set to Optimised with BatteryLife rather than Keep Batteries Charged, which can behave oddly at high SOC

On a larger three-phase system with multiple MPPTs and a grid meter like the VM-3P75CT, I imagine there are more places for signals to get confused — particularly around phase compensation and grid setpoint.

Has anyone else here dealt with grid draw persisting at high SOC on a multi-phase ESS? Curious whether it tends to be a comms/BMS issue or something deeper in the ESS logic itself.

@PaddyDavies odd one but have you checked whether the grid setpoint is drifting? I had something vaguely similar with my shepherd's hut setup — Victron Multiplus-C (single phase, nothing fancy) — where it kept pulling tiny amounts from the grid even with batteries sat at 100%. Turned out my ESS assistant had the feed-in threshold set just slightly off, so it was hunting around the setpoint rather than settling.

Also worth checking DVCC settings if you're running a Cerbo — I had mine interfering with charge current limits in a way that made no sense until I actually read the manual properly.

Not sure if any of this translates to three-phase but the ESS logic underneath is probably similar? Someone with actual Multiplus-II experience should probably weigh in though.

@PaddyDavies what firmware are you on? Victron pushed a few ESS updates last year that caused weird grid draw behaviour even with a full battery. Caught me out on the static van setup.

Also worth checking your DVCC settings — if shared voltage sense is off it can confuse the BMS and the Multi ends up pulling from grid as a kind of safety margin even when SOC looks fine.

Not a three-phase bloke myself but the logic should be similar. What's your grid setpoint sitting at? Anything above zero and it'll deliberately draw a bit regardless.

@PaddyDavies classic Victron "it's working as intended, honest" moment — check your minimum SOC setting in ESS, because if it's set even to 5% it'll pull from grid to maintain that buffer rather than drain the battery fully, even when the display looks like 100%.

Also worth poking at the DC feed-in toggle and your BMS assistant interaction — my Fogstar Drift cells threw similar nonsense until I realised the BMS was occasionally flagging a cell imbalance, Victron saw a "soft" high-voltage signal, and quietly decided grid was safer than pushing more charge through. Never logged it obviously, because that would be too helpful.

Check VRM portal logs at 1-minute resolution rather than the default aggregated view — the culprit usually hides in a 3-minute window that the hourly graph cheerfully smooths over entirely.

@PaddyDavies one thing nobody's mentioned yet — check your AC loads tab in the CCGX/Cerbo and see whether you've got any loads wired on the AC-input side rather than AC-output. ESS will pull from grid to feed those regardless of battery state, and it catches people out constantly. Had exactly this on my narrowboat when I rewired the inverter setup; shore power feeds were on the wrong side and the logs looked absolutely mental — full batteries, grid importing, no apparent reason.

Also worth pulling the VRM portal logs and cross-referencing grid draw timestamps against battery voltage curves. If the draw spikes are brief and irregular rather than sustained, you might be looking at a BMS communication dropout causing the ESS to go defensive rather than anything firmware-related.

Had almost the exact same head-scratcher when I wired up the shepherd's hut last spring — Victron ESS pulling from grid despite the Fogstar batteries showing well over 90%.

Turned out my grid setpoint was misconfigured. ESS tries to maintain a small grid feed-in or draw depending on that value — mine was sat at +100W rather than zero, so it was deliberately sipping from the meter constantly.

Worth checking yours in the ESS Assistant settings. Change it to 0W and see if the behaviour shifts. Also, if you've got any scheduled charging windows still active from initial setup, those can override what looks like perfectly sensible SOC logic and force a grid connection regardless.

@OffGridTel makes a good point about the AC loads tab too — definitely rule that out in parallel.